In the meantime I am trying to understand how the VCORE could work on the Soyo 5BT5, in my case there is a group of six jumpers (12 PIN), and in the photo there are visible the respective resistors, the values are of course all different, however it seems that there is a division, of the six total (two and four), the group of two is the one I imagine for voltages very close to 3.3V, the four should go down below 2.2V, for me it would be enough to even get to 2.0V, which for my K6-2+ 1.6V is more than tolerable, but let's see if we can figure out something else (?).

The resistances of the group of two are, R204 and R203, from what I read (we hope well) R204 is a 3093 and R203 is a 7153, from these numbers we will have to calculate the value in OHm 🕉.

The resistances of the group of four are, R202 R201 R200 and R248, and here I read, R202 is an 824, R201 is a 2103, R200 is an 1823, R248 is a 393, also here you have to calculate the value of each SMD resistance.

That R248 resistance would only be present in Rev.:2 motherboards. X, the rest is practically the same, so I would be of the opinion that an additional 393 resistance could lower the VCORE enough, but I don't think it goes down much below 2V.

I just got the results of the resistance values, let's see them:

R248 (393) corresponds to 39 KOHm

R200 (1823) are 182 KOHm

R201 (2103) are 210 KOHm

R202 (824) are 820 KOHm

R203 (7153?) I'm 715 KOHm

R204 (3093) are 309 KOHm

For the value of R203 I have the doubt that it is 2153 instead of 7153, unfortunately it does not read well, if it were 2153 then it would be 215 KOHm.

As you can see from the pages of the manual, what I thought was 2.2V is actually 2.1V, but we will check this with a voltmeter, in general the operation is through two jumpers, one for the I/O voltage (I think) and the other for the VCORE, for the VCORE there are four positions, while for I/O two.

The jumpers of my board are for single voltage 3.3V, because I use a Pentium 133 at the moment, but changing CPU with an MMX or a K6, I should automatically have the VCORE around 2.6V (because I have the jumper on 9-10 instead of 11-12), if it would actually be like that, I could put the jumper on 1-2 instead of 3-4 and theoretically I would get about 1.9V, which for me would be perfect

If instead I follow the manual and put the jumpers on 1-2 and 11-12, then 2.1V, but instead of the jumper I add a 39KOHm resistance, from what I see, I would get a higher voltage, maybe 2.3V or 2.4V, except that I remove the R248 resistance and put one of some KOHm.

This could be very useful if for example you want to use a 2.4V K6-2 or K6-IlI, with an additional resistance on the jumper, you could raise the VCORE from 2.1V to 2.4V, there are also some 2.45V Pentium MMX, which could go with the right VCORE voltage.

An alternative could be this, remove the R248 resistor, and weld a bridge, the voltage I had measured on PIN1 should be 1.65V, and this is fine for my K6-2+ 400, if you need to raise the one removed from R248 you could move it to the jumper, if you need the 2.1V, or change the value if you want for example 1.9V.

I did a boot test, first with the Pentium 133, later I changed CPU, and I put a Pentium 200 MMX, with this I made measurements, to look for VCORE lower than 2.8V, the PC always started, but the measurements on the Mosfet or in the inductor, were all quite strange, I have to try them again, because even changing position to the jumpers I see a strange 3.06V, even when it should only do 2.2V, I think I have to find the PINs in the back of the CPU (socket 7), and see what's really there, usually that type of measurement works very well, but Here maybe there are other components in the VCORE line, which could lower the voltage of 3.06V to the selected one.

This card does not have the VCORE self-detection function, and there is no BIOS function that can modify it, now I do not remember if there is a hardware monitor in the BIOS, with the relative voltages, I will see if there is and what it shows.

The Intel Pentium MMX 200 CPU, is of the blocked type, it only makes 166 and 200 MHz, so it is not possible to overclock at 233 and underclock at 133.

For now since the VCORE is high, I will not try any AMD K6-2 CXT, only when I am sure that the 2.2V corresponds to something similar, I will be able to do the CPU change, and redo the tests with undocumented VCORE settings.

The BIOS version is the factory one (?), it reads BT-1B2, and the motherboard should have a 1 Mbit BIOS chip, I will see later to update it to the BT-1B7 version, if I can make some progress with the undocumented VCORE.

I'm doing some tests on the Soyo SY-5BT, to see if there are undocumented VCORE settings, now everything seems quite realistic to me, maybe I did something wrong before...

First I try VCORE 2.8V, the two jumpers go on 3-4 and 11-12, on the Mosfet I measure 3.1V on the 1.8V jumpers.

Let's say that in absolute 3.1V are more than normal, later I will test to measure directly in the pins of socket 7, maybe there is a lower voltage?

Then I try VCORE 2.2V, moving the 3-4 to 1-2, and leaving the other on 11-12, on the Mosfet I have 2.33V and on the jumpers 1.08V, this is already good enough, but a little too high for the K6-2+, although we have to see what gets to the socket 7 pins?

With the previous selections, I move from 11-12 to 9-10, and strangely the measured values are practically the same, this means that the VIO probably changes, maybe it is lowered by 0.2V, but the VCORE is selected by the jumpers 1-2 3-4 5-6 7-8, while the VIO by the jumpers 9-10 11-12, so any attempt to be made is only on the jumpers from 1-2 to 7-8.

I tried to remove a jumper and leave only the one on 9-10, well on the Mosfet I measured 3.57V, and I stopped immediately, I don't know what was on the jumper, but it seems that without a jumper the voltage rises too much, maybe an extreme overclock would be possible, but I would leave it alone, overvolts like that would perhaps be useful if you want to make some records.

So I came to the conclusion, that probably the only way would be to try an additional resistance, I would expect the VCORE to go up, but maybe I'm wrong and go down?

I looked for a scheme of how the VCORE and the VIO work, I imagine it would be similar to the one I found, but it could very well be totally different, I'm trying to understand something, if there are any suggestions, I really appreciate them.

So taking for example another motherboard, I saw that it used a 39K resistance with a ground pin (black ATX wire), and actually it could be so that in the Soyo SY-5BT they made the jumper for the 2.2V, if so there should be the pin 2 connected to ground, we will check this as soon as possible, in the meantime following the logic of its calculations, I would have found which resistors to use to obtain voltages from 2.1V up to 1.6V, and they would be these:

2.1V R 22K or 27K | 2.0V R 18K

1.9V R 15K | 1.8V R 12K

1.7V R 10K | 1.6V R 8.2K

These are just calculations, and they need to be verified, so don't do anything and don't consider them correct, later I'll see if it actually works (?)

Then if I can get the lowest voltages that could be useful, I would switch to those between 2.2V and 2.8V, for these the resistors they used are 62K for 2.3V (K6-2 550), and 120K for 2.4V (K6-III and some K6-2), these resistors have been used successfully, the calculation is not my work.

After checking the pin 2, of the Soyo SY-5BT, I saw that it was not connected as I imagined, that is, it is not grounded, I did not waste time ⏱ to see where it goes, because for me it is useless to know, I would therefore have a solution to the problem, which could also be used on Rev.:1. X, on this you have to see if pins 2 and 4 (Rev.:2. X) they are connected together, if that's how it works for sure, to get the 2.2V you need a Resistance of 39KOHm (393), to be connected to pin 2 on the one hand, the other instead must be connected where the tracks converge and continue towards Q12, the idea is to skip the resistance of the individual jumpers, so that it goes to use the additional one (in this case 393 to have 2.2V), for this you will need a welding, which could be done with a piece of thin wire, in short I think it is the only fairly safe system, adding a resistance to the jumper, it would instead raise the VCORE, this solution I'm 99% sure it works.

The 1% is only due to my calculations, which could be optimistic (or pessimistic?), in the sense that I don't know if I would get the corresponding VCORE, or it would be a little lower, but I'm also sure of the values of 62KOHm for 2.3V and 120KOHm for 2.4V, although then we have to consider that the voltage measured on the Mosfet, could be slightly lower on the PINs of the Socket 7.

After all, if you have a minimum of skill, you can make a weld by uncovering a small part of the track, without having to stick in the solder of one of the resistors that go to the odd pins.

I continue with the testing of the Voltages on the Soyo SY-5BT, to be able to test the lower voltages, I need some resistors, unfortunately I have some missing, but I have to try four voltages, unfortunately I miss the 18K one that was necessary for the 2.0V, I will see to join two to obtain a value of just above 18K.

What resistances do I have?

I have a 10K for the VCORE 1.7V, this could be the starting one, but I'm thinking that maybe it's better to start with those closer to 2.2V, so I would start with a 22K or a 27K that are needed for 2.1V, of course theoretical, those measured could be a few tenths more.

I have two others, one 12K for VCORE 1.8V, and one 15K for 1.9V, these would theoretically not be of much use, but they would be useful for any stability problems in overclocking, the motherboard is equipped with FSBs greater than 66, and has if I understand correctly, two settings with FSB 75 and 83 with PCI at 33 MHz, this allows the installation of CPUs at 450 and 500 MHz, without problems on PCI cards, working at the normal frequency.

However, this may not work, in case we can settle for only 400 MHz, or 450 MHz with FSB 75, the PCI works at about 37 MHz, most VGAs will work, we will also have a higher RAM speed, in short, all in all if we do not find problems, that could be fine.

A little while ago I did a test, on the Soyo SY-5BT, it was practically as I thought, the even pins are connected to each other, at least those of the VCORE voltage, I didn't check those of the VIO, because I don't need to do anything, maybe it would be useful if I had a Pentium MMX 266 (Tillamok) between my paws, which has a VIO if I'm not mistaken, but many use it with VIO 3.3V, it probably heats up more but it works, however even having one (Kind gift from Santa Claus), I certainly wouldn't put it in this motherboard, which currently has 0KB of L2 cache, this Mod. VCORE is for AMD K6+ CPUs with VCORE from 1.6V to 2.0V, if the cache worked, I would have used a K6-2 CXT.

I connected the usual IDE 40PIN CF adapter with 256 MB and DOS (5.0) CF memory card, from QTPRO I saw that the FSB is not 66, but 68.5 (I had selected it finding this secret setting), so in reality the P200 goes to 206 MHz, a miserable gain, which however will be interesting using the K6-400, the real frequency will be 412 MHz, about 3% more, a modest increase that could give some frame/s in the various benches.

I started in recent days to do some tests on the Soyo SY-5BT, now I have reached the end, everything is very simple you need a resistor (the values are in the previous messages) to connect to the PIN2, and to the junction track of the R200-1-2 resistors.

For those who want to make the modification to the VCORE, I have highlighted in red the points where you can connect the resistor, you have to remove the jumper that you need to select 2.8V 2.9V and 3.2V, which is the one in correspondence with the R200 R201 and R202 resistors, you have to weld a wire or whatever leads, from the part where the resistors are joined by the track, looking at the image the bottom, a 39 KOHm resistor (2.2V VCORE) must be connected to the wire, this value was used at the factory in Rev.2. X, for those who have the Rev.1. X the minimum VCORE was 2.8V, but I think there are excellent chances that it can go down even beyond 2.0V, in the next few days I will try to make this mod., I will start by trying the 2.1V, using a 22K or 27K resistor, if I get about 2.1V, I will switch to other values, first from 1.9V and then from 1.8V and 1.7V.

Later I will also try the 2.0V by joining two resistors, to get to 18-19K, and I will do the same to get 1.6V with two resistors at about 8-9K total.

Wanting to do something more professional, you could build a mini circuit, in which to put all the desired VCORE, for example there are K6-2+ from 1.5V, if you want to use one of those, the resistance (I calculated but not verified) should be 7.5K OHm, so if I calculated well, you can go from 1.5V up to 2.4V, and make a jumper socket for each voltage, summaries are these:

1.5 1.6 1.7 1.8 1.9 2.0 2.1 2.2 2.3 2.4

Surely the value of 39K (2.2V) is exact, then if you measure 0.1 or 0.2 V more (blue arrts), it is likely that in the pins of Socket 7 there is instead the right voltage

I did some calculations, and the values of the resistances I calculated could be a little lower, I think it's better to start with a voltage higher than 2.2V, you can try to add resistors to the factory one (39K), the values are between 39K and 182K, theoretically 2.3V is obtained with a 62K resistor, and 2.4V with a 120K one, if this is the case for the values of 2.5V 2.6V and 2.7V we will be above 120K and below 182K, but you have to do some tests, and see if the voltages with 62K and 120k resistor correspond to 2.3V and 2.4V, so we will have to try a jumper on 2.2V with a resistor of 62-39= 23K, and with a resistor of 120-39= 81K, but on the market there are those of 22K and 82K, so I will use these.

Will it work? Yes of course, but will the VCORE be what results from the calculations I did? We'll see... I'll try the 22K resistor first, theoretically about 10K makes the VCORE go up by 0.1V, but to go up by 0.2V will 22K be enough?

However, all these theoretical speeches, in the end can be wrong (maybe not much?), the easiest thing, would be to try the individual voltages, changing the resistors, when I have all the values corresponding to the voltage, I would like to simply make a mod., remove the resistor from 2.2V and put one from 0 OHm, instead of the jumper I would put a cable that leads to an additional board, with a row of pins to place the jumper on the selected VCORE.

I've already done a first test with a jumper plus resistor, and it worked!

It was as I had imagined, that is, an additional resistor on the jumper, increases the VCORE, then we will see how much, but if we want to reduce it, we have to pass the SMD resistor solder on the MB in correspondence with the relative jumper.

Evidently, this circuit was made differently, compared to other MBs in which it is enough to add a resistor or a jumper, to obtain lower voltages, so any reference to these circuits, is useless, if for example we add another jumper, on the Soyo SY-5BT we will get higher VCORE voltages, maybe even too much, so be careful I absolutely do not recommend the use of a second jumper, just build a jumper with a resistor (as I did), to get the voltage you want, if you need an extra 0.05V, for greater stability, you can try Various values, it depends on where the jumper is positioned, if you want for example 2.95V, you can try with a 10K resistor and see if it's good, you'll probably need one greater than 10K, I don't know exactly, my calculations, are valid (?) For the lower frequencies.

Going back to my 5BT5, I remember that it is a Ver.:2. X, and which has an extra jumper to select the 2.2V, so after making a jumper with additional 10K resistor, I saw that the VCORE goes up from 2.36V to 2.47V, which means that I gained 0.1V with 10K, I can try to add other resistors, but just out of curiosity, in this way that is with the additional 10K, I would have selected 2.3V, although the reading on the tester says 2.47V, at the socket 7 pins there should be something less likely is 2.40V, and the difference of 0.1V could be due to the measuring instrument.

So to go below 2.2V, I would have to do a test with a resistor from less than 39K, not on the jumper though, but connected to the common track of the VCORE resistors and to the 2 pin.

This thing works both on Rev.:1. X that on Rev.:2. X, in any case you can try with the 10K resistor on the 2.8V and see the voltages before and after putting it on.

I wonder what minimum value (VCORE) is needed for an AMD K6-2+, that is, if you want to make a low frequency and low voltage PC, what VCORE do you need?

I know that there are versions for laptops, which if I'm not mistaken are 9W (or something like that), and I think my K6-2+ 400 1.6V is one of these, honestly I wouldn't need a voltage lower than 1.6V, but I think it works even with lower voltages, even if it may not be stable (?), however I got an idea, and so I would have decided that the minimum voltage of the VCORE that I could select, would be 1.3V, in this way it could start and work at frequencies below 200 MHz, why???

For a simple comparison with the Pentium MMX at the same frequency, you can also use it at 1.6V, but I would like to try it at lower VCORE, and do some bench, this only as a curiosity, because normally I will use it at just over 410 MHz (6X FSB 68.5).

During the last few days, I have done some calculations, to see which resistances to use to obtain the desired voltage, unfortunately for the lower voltages (see pdf MB), you need a mod., which basically can be done in different ways, a simpler one, solder a wire to the junction track of the resistors (which goes to Q12), and then connect it to the jumper + resistor, you can go up to the desired voltage by changing the value of the resistor. Another would be to remove the minimum voltage resistor and put one with a lower resistance, then go up with the voltage via jumper + resistor, always subtracting the value used, except in the case where we use a 0 OHm resistor 🕉

There would be others, one I would like to use is this one, weld a wire to the junction of the resistors, and build a simple circuit with all the desired voltages, you could go up from 1.3V up to 3.2V (always going up by 0.1V), then connect the other wire to pin 2 of the jumpers, for now the circuit is quite elementary, but nothing prevents you from making it easier by using multiple jumpers instead of just one, but I prefer it so in my opinion a single jumper avoids accidental errors, and limits a little the risk of selecting too high voltages, that's maybe it could be done A switch for lower voltages and higher voltages, in my case from 1.3V to 2.1V and from 2.3V to 3.2V.

I ran Phil's benches with the Soyo SY-5BT and P200MMX@206 and 0 KB cache L2, I started with the VCORE 2.2V, maybe I'm a little more than 2.3V actually, apart from that the PC seems to work well, so I proceed and start:

2) 105.2

The others 3) 4) and some others instead, they freeze or give error, I already know why, the VCORE is 2.8V while now I'm at 2.2V, I put it back to 2.8V and start again, now the system is stable, the scores are these:

3) 76.6 | 4) 32.7 | 5) 36.1 | 6) 15.7

A) 222.29 | b) 65.75 | c) 37.4 | d) 16.4

E) 14.8 | l) 1649.57 and 3452.30 | n) 153.74

At the moment, I don't have at hand 🖐, the values of the same benches that I did with the PCChips M550, but I imagine they are slightly higher, because in the SOYO I don't have a working L2 cache, the only thing is that I used a K6 233, but I think I also made benches with a Pentium MMX, so the comparison will be made between the two Pentium MMX.

Going back to the experimentation and verification of the relative voltages, on the SOYO 5BT5, I'm a little braked, because I'm missing various resistors, I don't have an assortment to choose what you need, however I've already verified that for the 2.3V you need a 473 (47K) resistor, and it must be replaced with that of the 2.2V which is a 393 (39K), instead for the 2.4V you need a 563 (56K) resistor, obviously we're not doing the precise ones, in the sense that if I measure 2.42V or 2.39V, they should be considered as 2.4V, in fact in the additional card I don't have half a tenth (0.05V).

But the fundamental reason is that on the market there are fixed values for resistors, so if in the case of 2.4V I get a value of 39+10+10=59K, in reality I have to use 56K instead, the difference is only a few cents, therefore acceptable.

For now I used a 10K resistor, and it corresponds to about 0.1V of increment, with two then 20K I have an increment of 0.2V, both go to add to the value of 39K, which would be the one present in the MB when you select the 2.2V.

To get to 2.5V, however, I think we need another 20K, because in the end we will get to about 180K, which is the 2.8V resistor.

I then checked the measured values, with a second voltmeter, this marks at least 0.05V less, it is probably more accurate, and considering that from the Mosfet to the pins there may be a further lowering, in the end on the CPU we will really have a little more than 2.2V, but the value is all in all correct, at least with voltages up to 2.4V, later we will see if it will be like this even with the lower ones, but for now I will limit myself to trying to reach 2.8V, for those lower than 2.2V, I think it will be necessary in addition to the mod., a change of CPU, I will start with a K6-2 CXT maybe a 350 or a 300, although the Pentium MMX seems to work even at relatively low voltages.

A little while ago I realized that there is an error in the drawing, the connection scheme was wrong, and since I wanted to make some adjustments, I made a new one.

First I eliminated the resistor to get 3.2V, because it already exists on the motherboard, and you can select by putting the jumper, even the 2.2V can be eliminated, because in my Rev.:2. X is present, so for those who have the 1. X must be added, so I preferred to move it to the bottom, but however other changes can be made, I am trying to understand if maybe they can be arranged differently to compact the space, in this board I will use traditional resistors (not SMD), in the future nothing prohibits using SMDs, to reduce the size a lot, currently they are about 10X5 cm, but you could eliminate a lot of PCB on the sides, obtaining an almost square shape, of about 6X5 cm.

In the next few days I will have to suspend the test, due to lack of resistors, maybe complete the research work for the VCORE 2.5V 2.6V 2.7V, I have to decide instead where to hook, to connect the junction of the resistors to the additional board, I hope to find a point where to solder, which is not dangerously close to something else.

Regarding the other works, I have some 486 boards in repair, a couple I would like to have them resurrected, in fact thinking about it there are at least three, each has its own particularity, one is ISA with UMC CPU (welded), apparently you can't see anything that can't be solved, but before trying it, I would like to clean some points, in the back there is a scratch that must be looked at under the microscope, maybe it didn't interrupt the tracks, if instead there are one or more interrupted points, they need to fix them.

According to the seller, the card is not working, but I don't know the level of competence what it is, it could surprisingly work, you have to try it and see what error codes it shows, let's hope well.

Another 486 motherboard, it has PCI slot and Socket 3, only 3 RAM 72 PIN, UMC chipset, this one is also not working according to the seller, I thought about how to clean the joints of some chips (including chipsets), they are a bit dark, instead of revivening them with the solder, I would like to give it a brush, but with something harder than a toothbrush, I will try something, I will try a corner, if the final result is acceptable, I will see if I continue, but only if I see the silver-colored joints back.

The last of these, is an ISA-OLB, OLB stands for Opti Local BUS, find any card of the kind, I believe at the limits of the impossible, however from what I know, alternatively you can use any ISA card, here unfortunately there is some repair to be done, the card has been slightly damaged by the battery, you need to remove and replace the socket of the keyboard chip (40 PIN), and the welds are quite hard, I suspended the work some time ago, because I would not want to damage it, however it would be possible to give it a try, or find some alternative solution, just to repair some traces, and see if Shows codes on the post card, currently I only see — —.

Going back to the SOYO 5BT5, I finished the experiment for voltages from 2.3V to 2.7V, I added various resistors, and I found values that can be quite correct, there is always a certain gap, due to both the precision of the voltmeter, and any voltage reductions, they are usually about 0.15V - 0.2V in total, so you measure 2.6V and they could be precisely 2.4V, I considered the factory ones good, and the gap increases, rising towards 3.0V, however I have not tried which resistors are used for 3.0V and 3.1V, for those I will use the K6-233 that I have on the M550, So by making a summary, and considering the commercial values of the resistors, we obtain this list:

2.2V 39K

2.3V 47K

2.4V 56K

2.5V 68K

2.6V 100K

2.7V 150K

Actually the 2.7V one was 140K, but on the market there are those of 150K or 120K, I chose the 150K, maybe it corresponds to 2.73V, but I think it's fine, also because that voltage is hardly used.

However, I am sure that it is possible to calculate these values, through a mathematical formula, unfortunately the one I used is not precise enough, so if someone is able to say what resistances to use for each individual voltage below 2.2V, it would prevent me from wasting time modifying the jumper, and I could start building the additional board for the VCORE, which I have already reduced and simplified enough, there is the possibility of reducing it further, using SMD resistors.

Today I go back to doing motherboard repairs, in the past I had done some work on an ASUS (which is a brand that I am very fond of), it is an S.462 A7V400-MX uATX format, with an AGP slot and three PCI, in the card there is an integrated VGA, in addition to the audio and the LAN, so it would be possible to use it without any additional card, the DDR RAM slots are only two, more than enough for such a card, also because I don't think it's possible to go beyond 2 GB, but it can very well be between 512 MB and 1 GB, a According to the Windows used.

What jobs have I done in the past?

I re-sold an SMD signed Q38, it was detached on one side, I repaired a track in the back, but I had interrupted further work due to the lack of another SMD signed D5, I tried to understand what the original was, I think it is a double diode (Sckottky), however I had not found a replacement, some time ago another user provided me with a photo of an identical board, which read "L43", I had not found anything between the scrap boards, at least until yesterday, looking for some resistors I saw an SMD signed L43!

Today I should be able to disconnect it, without damaging it and put it back on the ASUS, then we would have to add the heatsink on the NB, a RAM and a CPU, and try to start, who knows maybe it will give some signs or even start.?

The CPU that I will use initially, will be an AMD Duron, I think around 1000 MHz, later if necessary I could switch to something faster, ultimately it would not be the best, but it can be useful if installed in PC for Home - Office use.